When AI fails, insurance may pay – but who is at fault?
If an AI tool gives the wrong advice, who is actually responsible?
The company using it?
The developer behind it?
Or no one at all?
This is becoming a real question as tools like Claude and Copilot move into everyday work. Some companies are now insuring AI systems – covering the financial cost when things go wrong.
But insurance doesn’t answer the harder question: who is at fault?
I spoke to experts across security, strategy, and law to understand how this is actually being approached. The answer is less clear than you might expect.
